지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학술대회자료
- 저자정보
- 발행연도
- 2014.5
- 수록면
- 28 - 32 (5page)
이용수
초록· 키워드
오류제보하기
Distracted driving is a serious threat to traffic safety as evidenced by traffic accident records. Detection of driver distraction is not straightforward due to the complex nature of driver behaviors. In this paper, we develop a cloud model to detect driver visual distraction effectively and efficiently. The cloud model is a new mathematical representation of linguistic concepts, which implements uncertainty transformation between qualitative concepts and their quantitative values based on mathematical statistics and fuzzy mathematics. The model has three numerical characteristics: Expected Value_ Ex, Entropy_ En and Hyper-entropy_ He. The Ex represents the most suitable point to represent the domain of a concept which has been quantified. The En represents the uncertainty measurement of qualitative concept, reflecting the range of domain space: the larger the entropy, the larger the granularity, the concept is more macro). The He represents the uncertainty measurement of entropy, representing the relationship between randomness and fuzziness. We conducted a driving simulator experiment to obtain data for distracted driving based on manipulation of a navigation system during driving. We chose the standard deviation of steering wheel angles and lateral lane position as the distraction indicators. We drew the receiver operating characteristic curves of the two variables to confirm their effectiveness for detecting driver distraction, and also to determine threshold values for the cloud model. We trained the cloud model using the simulator experiment results. We then tested the cloud model using the new simulator experiment results. The results show that the cloud model can detect driver visual distraction effectively and efficiently because the model extracts three numerical characteristics from one variable, and therefore enables us to use fewer variables for reliable detection while avoiding possible conflicts among many variables.
목차
ABSTRACT
1. Introduction
2. Method
3. Results
4. Conclusion
References
참고문헌 (0)
참고문헌 신청이 논문의 저자 정보
최근 본 자료
댓글(0)
0
UCI(KEPA) : I410-ECN-0101-2016-530-002491436